Configure Windows Server Failover Clustering

Windows Server Failover Clustering (WSFC) provides high availability for critical workloads by automatically failing over services between cluster nodes when hardware or software failures occur. This guide covers setting up a two-node failover cluster on Windows Server VPS instances.

Prerequisites

• Two or more Windows Server VPS instances with identical configurations
• Active Directory: All nodes must be domain-joined
• Shared storage: iSCSI SAN, Storage Spaces Direct, or shared disk
• Multiple NICs: Separate networks for cluster heartbeat and client traffic
• Same Windows Server version on all nodes

Install Failover Clustering Feature

# Install on ALL cluster nodes
Install-WindowsFeature -Name Failover-Clustering -IncludeManagementTools

# Verify
Get-WindowsFeature Failover-Clustering

Validate Cluster Configuration

# Run validation tests before creating the cluster
# This checks network, storage, and system configuration
Test-Cluster -Node "node1.corp.example.com", "node2.corp.example.com" `
    -Include "System Configuration", "Network", "Storage"

# Review the validation report
# Located at: C:\Users\Administrator\AppData\Local\Temp\Validation Report*.html

# For VPS environments where shared storage may differ
Test-Cluster -Node "node1", "node2" -Include "System Configuration", "Network"

Create the Cluster

# Create a new failover cluster
New-Cluster -Name "MyCluster" `
    -Node "node1.corp.example.com", "node2.corp.example.com" `
    -StaticAddress 10.0.0.100 `
    -NoStorage

# Verify cluster is running
Get-Cluster
Get-ClusterNode | Format-Table Name, State, NodeWeight

# Configure cluster quorum
# For 2-node clusters, use a file share witness or cloud witness

# File Share Witness
Set-ClusterQuorum -NodeAndFileShareMajority "\\fileserver\ClusterWitness"

# Cloud Witness (Azure Storage Account)
Set-ClusterQuorum -CloudWitness `
    -AccountName "mystorageaccount" `
    -AccessKey "YourAzureStorageKey"

Configure Cluster Networks

# View cluster networks
Get-ClusterNetwork | Format-Table Name, State, Role

# Rename networks for clarity
(Get-ClusterNetwork "Cluster Network 1").Name = "Client Network"
(Get-ClusterNetwork "Cluster Network 2").Name = "Heartbeat Network"

# Set network roles
# 0 = Not used by cluster
# 1 = Cluster only (heartbeat)
# 3 = Client and cluster
(Get-ClusterNetwork "Client Network").Role = 3
(Get-ClusterNetwork "Heartbeat Network").Role = 1

Add Clustered Roles

# Generic Service (e.g., for a custom application)
Add-ClusterGenericServiceRole -ServiceName "MyAppService" `
    -Name "MyApp" `
    -StaticAddress 10.0.0.101

# File Server Role
Add-ClusterFileServerRole -Name "ClusterFS" `
    -Storage "Cluster Disk 1" `
    -StaticAddress 10.0.0.102

# SQL Server Always On (after SQL is installed on both nodes)
# Enable Always On in SQL Server
Enable-SqlAlwaysOn -ServerInstance "node1" -Force
Enable-SqlAlwaysOn -ServerInstance "node2" -Force

# Hyper-V Virtual Machine (clustered VM)
Add-ClusterVirtualMachineRole -VMName "ImportantVM"

Test Failover

# Manual failover for testing
Move-ClusterGroup "MyApp" -Node "node2"

# Verify the role moved
Get-ClusterGroup "MyApp" | Select-Object Name, OwnerNode, State

# Move back
Move-ClusterGroup "MyApp" -Node "node1"

# Simulate node failure (careful in production!)
# Stop-ClusterNode "node1"  # This will trigger automatic failover

# Check failover events
Get-ClusterLog -Destination C:\ClusterLogs -TimeSpan 60

Monitoring

# Check cluster health
Get-ClusterNode | Select-Object Name, State
Get-ClusterGroup | Select-Object Name, OwnerNode, State
Get-ClusterResource | Select-Object Name, OwnerGroup, State

# Cluster-aware updating (rolling patches)
$cau = New-CauOrchestrator
Invoke-CauRun -ClusterName "MyCluster" `
    -MaxRetriesPerNode 3 `
    -RequireAllNodesOnline `
    -EnableFirewallRules `
    -Force

# Cluster event monitoring
Get-WinEvent -LogName "Microsoft-Windows-FailoverClustering/Operational" `
    -MaxEvents 20 | Format-Table TimeCreated, Message

Storage Spaces Direct (S2D) for VPS

# Enable Storage Spaces Direct (Windows Server 2022 Datacenter)
# Each node contributes local storage to a shared pool

Enable-ClusterStorageSpacesDirect -PoolFriendlyName "S2D Pool" `
    -Confirm:$false

# Create a virtual disk
New-Volume -FriendlyName "ClusterVolume1" `
    -FileSystem CSVFS_ReFS `
    -StoragePoolFriendlyName "S2D Pool" `
    -Size 100GB `
    -ResiliencySettingName Mirror

Best Practices

• Always run validation tests before creating or modifying a cluster
• Configure a witness: Essential for 2-node clusters to prevent split-brain
• Separate heartbeat network: Use a dedicated network for cluster communication
• Test failover regularly during maintenance windows to ensure it works
• Use Cluster-Aware Updating for rolling patches without downtime
• Monitor cluster health and set up alerts for node failures
• Document your cluster configuration including network, storage, and role assignments